can you share the result of the below query ?

select count(1) from transaction_attributes where analysis_start_date = '2025-05-01' and analysis_end_date = '2025-05-01' , 

If it has multiple entries , the join condition will lead to cross join and hence the spark job is running for ever